Peer review is a system built on trust, and when that trust is broken by an AI-generated review that masquerades as human critique, the entire process suffers. The experience described here, a weak rejection with high confidence from one reviewer, while four others offer positive, low-confidence scores, is not an isolated anomaly. It is a symptom of a growing problem: the use of large language models to fabricate feedback that sounds plausible but lacks the substance of genuine engagement.
The practical question for any researcher facing this is not whether to act, but how to act with precision. Collecting evidence is the first step, and you already have it. The trivial points, the irrelevant baselines, the identical phrasing to your own LLM simulations, these are not coincidences. Document them systematically. Save the review text, your rebuttal, and the silence that followed. Compare the reviewer's language patterns against known LLM outputs, and note the absence of any response to your substantive counterpoints. This is not about proving intent; it is about demonstrating inconsistency, and that is a standard an area chair can understand.
When you bring this to the AC, focus on the quality of the review, not the accusation of LLM usage. The latter is difficult to prove and may feel like a personal attack. The former is objective: the review is shallow, misaligned with your paper's scope, and unresponsive to rebuttal. Frame it as a request for procedural fairness, this reviewer did not engage with your work, and their high-confidence rejection is unsupported by the evidence you presented. The AC's job is to weigh reviewer signals, and a pattern of disengagement is a legitimate signal to flag.
What this means for you, practically, is that silence is not a strategy. Submit your rebuttal, then follow up with the AC directly. Share the evidence you have collected, and ask for their guidance on how to handle a reviewer who appears to have abandoned the process. Many ACs will appreciate the transparency, because they too are struggling to filter out low-quality reviews that erode confidence in the system. You are not tattling; you are advocating for the integrity of your own work, and that is a reasonable, professional move.
The broader lesson is that AI-generated reviews will only increase, and researchers need a playbook for responding that is both firm and fair. We do not advocate for witch hunts, but we do advocate for accountability. If a review cannot withstand a simple test of relevance and engagement, it should not carry the same weight as one that does. Your evidence is your leverage, and your clarity is your strength. Use both, and you will not only protect your paper, you will help set a standard for how the community handles this new, unwelcome reality.